The generalized stable equivalence problem (Q1952156)

Let \(X\) be an affine variety over a field \(k\). Two hypersurfaces \(H_1, H_2\subseteq X\) are \textit{equivalent (in \(X\))} if there exists an automorphism \(f\) of \(X\) mapping \(H_1\) onto \(H_2\). They are \textit{stably equivalent} if there exists \(n\) such that \(H_1\times \mathbb{A}^n_k\) and \(H_2\times \mathbb{A}^n_k\) are equivalent in \(X\times \mathbb{A}^n_k\). The generalized stable equivalence problem for \(X\) asks whether any two stably equivalent hypersurfaces in \(X\) are actually equivalent. The problem was studied mostly for \(X=\mathbb{A}^n_k\), where for \(k\) algebraically closed of characteristic zero an affirmative answer has been given if one hypersurface is a zero set of a so-called \textit{test polynomial for monomorphism} [\textit{V. Shpilrain} and \textit{J.-T. Yu}, J. Algebra 251, No. 1, 295--307 (2002; Zbl 1067.14064)] (a notion introduced by \textit{Z. Jelonek} [J. Pure Appl. Algebra 147, No. 2, 125--132 (2000; Zbl 0972.14004)]) or if one hypersurface is not uniruled [\textit{R. Dryło}, Ann. Pol. Math. 87, 93--98 (2005; Zbl 1090.14019)]. For \(X=\mathbb{A}^2_k\) and an arbitrary field \(k\) the problem has been solved positively in [\textit{L. Makar-Limanov, P. van Rossum, V. Shpilrain} and \textit{J.-T. Yu}, Comment. Math. Helv. 79, No. 2, 341--349 (2004; Zbl 1063.14077)] (case \(\text{char}k=0\)) and by the author in [\textit{R. Dryło}, Colloq. Math. 124, No. 2, 247--253 (2011; Zbl 1227.13004)]. In general the problem remains open for \(X=\mathbb{A}^n_k\), \(n\geq 3\). The well-known example, due to Danielewski, of non-isomorphic surfaces in \(\mathbb{C}^3\) with isomorphic cylinders was used in Proposition 1.7 of [\textit{V. Shpilrain} and \textit{J.-T. Yu}, J. Algebra 251, No. 1, 295--307 (2002; Zbl 1067.14064)] to show that if we put no additional assumptions on \(X\) then the answer to the stable equivalence problem is negative (in this example \(H_1\) and \(H_2\) are non-isomorphic). In this article the author shows (Theorem 1) that if \(X\) is not \(k\)-uniruled (for \(k\) uncountable this means that \(X\) is not dominated by \(Y\times \mathbb{A}^1_k\), where \(Y\) is an affine variety of dimension \(\text{dim}X-1\)) then the answer is positive. The case \(\text{char}k=0\) was already done in [\textit{R. Dryło}, Ann. Pol. Math. 92, No. 1, 41--48 (2007; Zbl 1128.14042)]. The author shows also (Proposition 2), based again on the idea by Danielewski, that gluing two cylinders over a nonrational affine curve one can construct a smooth affine surface \(X\) containing two stably equivalent lines which are not equivalent in \(X\).